Research On Microwave Tumor Detection Technology Based On Genetic Algorithms

Breast cancer is one of the major diseases that threaten women's health.It is especially important for the detection of breast tumors,especially early tumors.In recent years,microwave tumor detection technology has become a hot topic at home and abroad.Compared with conventional breast cancer detection technology,microwave tumor detection technology has the characteristics of small radiation,low cost and high accuracy.According to the significant difference of electromagnetic parameters between tumor tissue and normal tissue,microwave tumor detection technology is divided into active microwave technology,passive microwave technology and microwave ultrasonic imaging,while microwave tomography imaging(MTI)in active microwave technology is the main method for detecting tumors by microwave research direction.The essence is to extract the electromagnetic signal around the tissue through the near-field radiation of the antenna array,deduct the electromagnetic parameter distribution inside the tissue,and perform the parameter imaging to obtain the tumor information.The use of tissue peripheral electromagnetic signals to deduct the electromagnetic parameters inside the tissue is an electromagnetic inverse scattering study.It is necessary to solve the nonlinear problem of inverse scattering by combining the optimization algorithm.In response to this problem,this thesis introduces artificial intelligence genetic algorithm(GA),combined with microwave tomography to achieve breast tumor detection.In view of the fact that XFdtd electromagnetic simulation software can only provide 64 kinds of materials and the computational complexity of genetic algorithm,the research on early tumor microwave detection evolves from low resolution to high resolution,mainly to solve breast modeling,parameter setting,data integration,and inheritance.Iterative and other issues,the specific content is as follows.Firstly,the time domain finite difference method(FDTD)and the electromagnetic simulation software XFdtd based on this algorithm are briefly introduced,and the simulation model is established.Then the basic principle of GA is introduced,and the way of GA to realize tumor microwave detection is demonstrated.Due to the large amount of GA calculation and long iteration time,this paper introduces the confocal microwave imaging(CMI)technology in active microwave technology to find the approximate location of the tumor and reduce the initial imaging area.Then the microwave tomography based on standard genetic algorithm(SGA)is adopted.Tumor detection was performed in the reduced imaging area,and the resolution was gradually increased from 5 mm to 2 mm.However,SGA has the defects of premature convergence and slow convergence.Moreover,the SGA-based microwave tomography imaging technique can only improve the resolution to 2mm when detecting tumors with a diameter of 10 mm.To this end,this paper proposes an improved genetic algorithm(IGA)tomographic imaging technique,which adaptively modifies the mutation probability according to the diversity of the population and the fitness value of the individual during the genetic iteration.In the case of loss of diversity,the probability of individual variation with a large fitness value is reduced,and the probability of individual variation with a small fitness value is increased.The simulation results show that IGA can effectively avoid premature convergence and improve the convergence speed.The proposed method can achieve 1mm resolution tumor detection when detecting tumors with a diameter of 4mm,and the imaging accuracy is as high as 91.7%.